Vitreous fluorophotometry: identification of sources of variability.

Slit-lamp vitreous fluorophotometry (VF) has been used by several investigators to document the early breakdown of the blood-ocular barrier. After an injection of fluorescein, levels of fluorescein in the vitreous humor are reported to be elevated in diabetic patients compared with controls even when there is no retinal pathologic condition discernible by angiography or ophthalmoscopy. We have examined the technique of VF in normal persons and have identified several potential sources of error in the determination of vitreous fluorescein levels. Fluorescein in the choroidal circulation significantly influences VF readings taken at the retinal surface and in the posterior vitreous. This effect is modulated by ocular pigmentation, with heavy pigmentation of dark-eyed normals more effectively screening out choroidal fluorescence than the light pigmentation of blue-eyed subjects. Influence of choroidal fluorescein extends toward the midvitreous, and the range of influence is directly related to the slit width of the exciting light. At narrow slit widths the effect is diminished but can never be completely eliminated. Blood fluorescein levels decrease with time, and since choroidal fluorescein contributes to VF values, the time after injection that VF reading are taken is a critical parameter.